July 23, 2026, (Inside AI) — Anthropic, the maker of Claude, is weighing mandatory 10b5-1 trading plans for all employees after its expected IPO, according to a report by The Information citing people familiar with the matter. The move aims to preempt insider-trading risks by requiring rank-and-file staff to sell shares only through preset schedules, a departure from the typical voluntary use of such plans by executives.
The deliberation signals how deeply AI startups are scrutinizing governance as they hurtle toward public markets. With valuations surging—Anthropic was last pegged at $61.5 billion—and regulators eyeing the sector, the company appears to be erecting guardrails early. OpenAI, its chief rival, is also racing to go public, intensifying pressure to set standards for frontier AI firms.
10b5-1 plans allow insiders to trade stock at predetermined times and prices, providing an affirmative defense against insider-trading allegations. While common among C-suite officers, extending them to all employees is rare and logistically complex. It would require careful coordination with lock-up expirations and blackout periods, potentially limiting liquidity for staff accustomed to trading freely.
Anthropic’s structure adds another layer. Founded as a public benefit corporation with a long-term trust mechanism, it has emphasized safety over speed. Mandatory trading plans align with that ethos but could also be seen as a recruiting hurdle in a talent war where equity liquidity is a prized perk.
The report did not detail how such plans would be implemented across different jurisdictions or for former employees. Legal experts note that while 10b5-1 plans offer protection, they are not a shield if the plan was entered into while possessing material non-public information. A 2023 SEC rule change also introduced a cooling-off period for executives, but it's unclear how that would apply to broad-based employee plans.
Anthropic declined to comment. The company has not publicly confirmed IPO timing, though filings could come as early as 2026. Its revenue reportedly hit $1 billion annually, fueled by enterprise adoption of Claude and partnerships with Amazon and Google.
The insider-trading spotlight on AI firms isn't new. In 2024, the SEC charged a former Palantir employee with trading on confidential data about an AI contract. As AI companies amass sensitive information—from model performance metrics to government deals—the scope for misuse widens.
Anthropic’s move, if finalized, could set a precedent for other AI startups eyeing public debuts. It also underscores a maturing industry grappling with the less glamorous mechanics of being a public company.
